Alfresco just released their source under GPL with a FLOSS exception.
This move is about choice, in other words. Choice for our customers (Freedom of access to the world’s best content management system). Choice for our partners. Choice for our competitors (join us or get run over
. Choice for the community.
…The GPL is open source’s best business license. It is open source capitalism – the free market at its finest.
Freedom pays great dividends.
…the proprietary software companies we compete with have a limited shelf life.
Welcome Alfresco and kudos. Enough of this non-OSI-approved modified MPL bullshit. If your license is not approved by OSI then you’re not open source. Real simple. Worse yet you have companies like Socialtext. I think I’ve remained silent about these guys too long. Socialtext has claimed to be open source and fervently waved this banner since their founding in 2002. However, they didn’t release a stitch of source code until July, 2006. Four years later. WTF? When they did release their source they did so under a non-OSI-approved modified MPL license. WTF? Most recently Mayfield suggested that because they’ve allegedly "supported Kwiki for years" this makes Socialtext open source. I’ve been using Kwiki at every O’Reilly event (I love O’Reilly) and to the best of my knowledge Kwiki hasn’t changed in three years! I’ve kept my mouth shut because they’re competitors and I assumed others would have cried foul over their behavior long ago. Oddly, no one has. Then a few months ago (around Thanksgiving 2006) murmuring started about attribution licenses, which really is of lesser concern as far as Socialtext is concerned. Berlind blogged about how he can see it either way blah blah. In the end he called for "disclosure" as to whether the license is OSI approved or not. If you’re not, you’re not open source! What’s there to disclose? I responded here. Shortly thereafter Berlind sent me an email titled: "Laughable." It may have been a tad condescending, but in general he stated that disclosure is a good thing he didn’t see how anyone could disagree. Sure, I was a tad harsh, but I didn’t get the impression he was saying anything of use and I was also a tad offended by his statement: "the supposed keeper of the official definition of ‘open source’ and the consortium to which open source license authors typically turn to have their licenses ratified as adhering to that definition". Supposed? Typically turn? Read his post, read my response, you be the judge. As a side note, isn’t Berlind on the advisory board for Socialtext? I know one of those ZDNet blogger dudes is and I thought it was him. I’m not certain. UPDATE: I don’t think it’s Berlind, but I know there is a ZDNet journalist/blogger guy who is on Socialtext’s advisory board. Update’: My bad, it’s Mitch Ratcliffe.
Nat Torkington while organizing OSCON (which I love and is where we launched www.opengarden.org, OSCON06) recently asked: "Is ‘Open Source’ Now Completely Meaningless?" Well, if we continue down this road I don’t see how it couldn’t become meaningless. By the way Nat, I don’t think you should make a hard and fast rule as to whether you should disallow closed source companies to participate in OSCON. Do it on a case by case basis because surely there will be some worthy exceptions. Anyway, back on subject. What I propose is that OSI manages a wall of shame for companies that behave inappropriately with respect to use of the term "open source". It could work something like this. 1). The accused company receives a public warning via email and it’s also published online (posted on a wiki perhaps). Along with the warming the company is asked to cease their inappropriate use of the term open source and provided clear steps for complying with OSI. 2). The accused follows the steps for compliance in the provided timframe or they receive public censure, which starts with a public admonishment of their actions and could be escalated. All this is archived (hence the wiki suggestion) and indexed. Perhaps OSI could work with SourceForge.net and other sites to create a coalition that could perfrom the censure.
Why is this better than the current process? Well right now it’s very ad hoc. Only geeks know what’s going on because we’re the only ones willing to participate in a rabid discussion list. Ultimately OSI has no stick to wield and everything is pretty quiet outside geek circles. Even within geeky circles people are confused and it’s unclear if someone is violating the will of OSI. Also, this way when someone performs a web search for the aformentioned accused company the warning and censure would be found in the search result set. Thus providing a monetary incentive for compliance with OSI’s will. I very recently emailed these thoughts (mostly) to Tiemann. UPDATE: Michael and I spoke. I talk about it in this blog post. Something needs to be corrected though because this is a growing trend that seems to be spurred on by the flood of cash resulting from the venture capitalists’ interest in open source. And to compound the problem you have very media savvy folks with deep deep pockets that are very clearly manipulating public perception by injecting their will into journalism (mostly bloggers).
Anyway, enough soap-boxing. Back to Alfresco. What’s the intent of the Alfresco FLOSS exception (also employed by MySQL and others)? It’s meant, as far as I can tell, to maximize freedom in extending and integrating with dissimilarly licensed FLOSS software. Meaning, whatever you extend or integrate can continue to maintain it’s own licensing as long as it’s OSI approved (or on a list of OSI approved licenses that Alfresco provides). We achieved the same end result at MindTouch by providing DREAM under LGPL. DREAM is our Distributed REST Application Manager and what we’re building MindTouch DekiWiki on top of. In fact, we’re slowly discarding MediaWiki PHP logic for C# on Mono/Net 2.0 in the form of DREAM services. And as you would imagine DREAM is also powering our API. It will should be technically feasible that ultimately one could install just a PHP layer on their shared server and have a very sophisticated Service Oriented Distributed Architecture (SODA) powered by DREAM providing all the business logic. Think about that for a moment. It has huge benefits and enormous ramifications. Anyway enough about us. Is this FLOSS exception OSI approved? I saw one reference to it being an OSI approved exception, but I really don’t know if it is. It’s exception is only for other OSI approved licenses, it’s got to be.
